Azure的AI使用-(语言检测、图像分析、图像文本识别)

1.语言检测

安装包：

# 语言检测
%pip install azure-ai-textanalytics==5.2.0

需要用到密钥和资源的终结点，所以去Azure上创建资源，我这个是创建好的了然后点击密钥和终结者去拿到key和终结点

两个密钥选择哪个都行 

语言检测代码示例： 

key = ""
endpoint = ""
from azure.ai.textanalytics import TextAnalyticsClient
from azure.core.credentials import AzureKeyCredential

def authenticate_client():
    ta_credential= AzureKeyCredential(key)
    text_analytics_client=TextAnalyticsClient(
        endpoint=endpoint,
        credential=ta_credential
    )
    return text_analytics_client
    
client=authenticate_client();

# 检测文本是哪种语言
def language_detection_example():
    try:
        documents = ["Ce document est rédigé en Français."]
        response=client.detect_language(documents=documents,country_hint = 'us')[0]
        print("response",response)
        print("Language: ", response.primary_language.name)

    except Exception as err:
        print("Encountered exception. {}".format(err))
language_detection_example()

运行结果：

response {'id': '0', 'primary_language': DetectedLanguage(name=French, iso6391_name=fr, confidence_score=1.0), 'warnings': [], 'statistics': None, 'is_error': False, 'kind': 'LanguageDetection'}
Language:  French

2.提取关键短语

# 提取关键语言
def key_phrase_extraction_example(client):

    try:
        documents = ["你好啊，我叫feng，是java程序员，想学习更多的知识"]

        response = client.extract_key_phrases(documents = documents)[0]

        if not response.is_error:
            print("\tKey Phrases:")
            for phrase in response.key_phrases:
                print("\t\t", phrase)
        else:
            print(response.id, response.error)

    except Exception as err:
        print("Encountered exception. {}".format(err))

key_phrase_extraction_example(client)

 返回：感觉对中文的提取一般不是很友好

Key Phrases:
		 feng
		 java程
		 你好
		 想学
		 多的知识

换成英文 

documents = ["Dr. Smith has a very modern medical office, and she has great staff."]

 关键字提取好像就会好很多啊！

 ["Hello, my name is Feng. My hobby is singing and traveling, and I hope to make friends with you"]

确实英语就好很多。 

 

2.图像分析

安装包：

# 图像分析
%pip install --upgrade azure-cognitiveservices-vision-computervision
# 图像处理库
%pip install pillow

这是3.2版本的，这个版本可以支持返回分析中国语言

它就是给一个图片，它会分析出图片大概有什么，以及占的比例，就像是百度的识别万物一样，识别出的物品是什么以及占比。

2.1 url图片地址分析-版本3.2

咱们拿这个图片让它帮分析一下

代码示例： 

# 图像分析-url版本
from azure.cognitiveservices.vision.computervision import ComputerVisionClient
from azure.cognitiveservices.vision.computervision.models import OperationStatusCodes
from azure.cognitiveservices.vision.computervision.models import VisualFeatureTypes
from msrest.authentication import CognitiveServicesCredentials
import os

os.environ["VISION_KEY"]=''
os.environ["VISION_ENDPOINT"]=''
subscription_key = os.environ["VISION_KEY"]
endpoint = os.environ["VISION_ENDPOINT"]
computervision_client = ComputerVisionClient(endpoint, CognitiveServicesCredentials(subscription_key))

remote_image_url = "https://raw.githubusercontent.com/Azure-Samples/cognitive-services-sample-data-files/master/ComputerVision/Images/landmark.jpg"
tags_result_remote = computervision_client.tag_image(remote_image_url,language="zh")
if (len(tags_result_remote.tags) == 0):
    print("No tags detected.")
else:
    for tag in tags_result_remote.tags:
        print("'{}' with confidence {:.2f}%".format(tag.name, tag.confidence * 100))
print()

运行结果：

 2.2 本地图片分析-版本3.2

# 图像分析-本地图片
from azure.cognitiveservices.vision.computervision import ComputerVisionClient
from azure.cognitiveservices.vision.computervision.models import OperationStatusCodes
from azure.cognitiveservices.vision.computervision.models import VisualFeatureTypes
from msrest.authentication import CognitiveServicesCredentials
import os

os.environ["VISION_KEY"]=''
os.environ["VISION_ENDPOINT"]=''
subscription_key = os.environ["VISION_KEY"]
endpoint = os.environ["VISION_ENDPOINT"]
computervision_client = ComputerVisionClient(endpoint, CognitiveServicesCredentials(subscription_key))

local_image_path = os.path.join("C:\\Users\\Uniigym3\\AppData\\Roaming\\Python\\Python38\\Scripts\\images", "11.png")
local_image = open(local_image_path, "rb")
tags_result_local_image = computervision_client.analyze_image_in_stream(local_image,language="zh")
print(tags_result_local_image)
if (len(tags_result_local_image.categories) == 0):
     print("No description detected.")
else:
    for category  in tags_result_local_image.categories:
        print("'{}' with confidence {:.2f}%".format(category.name, category.score  * 100))
print()

运行结果：

 2.3 url图片地址分析-版本4.0

安装包：

%pip install azure-ai-vision

咱们让它分析这个图片

import os
import azure.ai.vision as sdk

service_options = sdk.VisionServiceOptions("",
                                           "")

vision_source = sdk.VisionSource(
    url="https://learn.microsoft.com/azure/ai-services/computer-vision/media/quickstarts/presentation.png")
analysis_options = sdk.ImageAnalysisOptions()
# 可选的视觉特征
analysis_options.features = (
    sdk.ImageAnalysisFeature.CAPTION |
    sdk.ImageAnalysisFeature.TEXT
)
analysis_options.language = "en"
# 性别中立的描述文字,默认值为区分性别的描述文字。 例如，在英语中，当你选择性别中立的描述文字时，“女性”或“男性”等术语将替换为“人员”，而“男孩”或“女孩”则将替换为“儿童”。
analysis_options.gender_neutral_caption = False
image_analyzer = sdk.ImageAnalyzer(service_options, vision_source, analysis_options)
result = image_analyzer.analyze()
# 成功你就按自己选的特征进行
if result.reason == sdk.ImageAnalysisResultReason.ANALYZED:
    if result.caption is not None:
        print(" Caption:")
        print("   '{}', Confidence {:.4f}".format(result.caption.content, result.caption.confidence))
    if result.text is not None:
        print(" Text:")
        for line in result.text.lines:
            points_string = "{" + ", ".join([str(int(point)) for point in line.bounding_polygon]) + "}"
            print("   Line: '{}', Bounding polygon {}".format(line.content, points_string))

else:
     error_details = sdk.ImageAnalysisErrorDetails.from_result(result)
     print("   Error reason: {}".format(error_details.reason))
     print("   Error code: {}".format(error_details.error_code))
     print("   Error message: {}".format(error_details.message))

运行结果：除图片的信息展示以外还会反馈出图片的文字 

analysis_options.gender_neutral_caption = True ,性别中立的描述文字,默认值为区分性别的描述文字。 例如，在英语中，当你选择性别中立的描述文字时，“女性”或“男性”等术语将替换为“人员”，而“男孩”或“女孩”则将替换为“儿童”。

如果设置False或不加这个设置，刚才的结果就是

Caption:
   'a man pointing at a screen', Confidence 0.7768

2.4 本地图片分析-版本4.0

就只需要把上面的这个url图片的代码改成下面的图片路径代码就可以直接在本地使用了。

vision_source = sdk.VisionSource(url="https://learn.microsoft.com/azure/ai-services/computer-vision/media/quickstarts/presentation.png")

vision_source = sdk.VisionSource(filename="C:\\Users\\Uniigym3\\AppData\\Roaming\\Python\\Python38\\Scripts\\images\\test.jpg")

我们测试个百变小樱魔术卡

运行结果：

说是有卡通的小女孩，并且标签也识别出日本动漫。 

再来测试个图片：好几个国家的语言哈

 运行结果：都能轻松的识别出来

官网图片示例：多种图片https://github.com/Azure-Samples/cognitive-services-sample-data-files/tree/master/ComputerVision/Images

图像分析3.2版本git示例：https://github.com/Azure-Samples/cognitive-services-quickstart-code/blob/master/python/ComputerVision/ImageAnalysisQuickstart.py

3.图像OCR文本识别

3.1 url图像地址识别

用这个图片来测试下

#OCR文本识别
from azure.cognitiveservices.vision.computervision import ComputerVisionClient
from azure.cognitiveservices.vision.computervision.models import OperationStatusCodes
from azure.cognitiveservices.vision.computervision.models import VisualFeatureTypes
from msrest.authentication import CognitiveServicesCredentials
import time

computervision_client = ComputerVisionClient(You endpoint, 
                                             CognitiveServicesCredentials(Your key))
read_image_url = "https://learn.microsoft.com/azure/ai-services/computer-vision/media/quickstarts/presentation.png"
read_response = computervision_client.read(read_image_url,  raw=True)
read_operation_location = read_response.headers["Operation-Location"]
operation_id = read_operation_location.split("/")[-1]
while True:
    read_result = computervision_client.get_read_result(operation_id)
    if read_result.status not in ['notStarted', 'running']:
        break
    time.sleep(1)
if read_result.status == OperationStatusCodes.succeeded:
    for text_result in read_result.analyze_result.read_results:
       for line in text_result.lines:
           print(line.text)
           print(line.bounding_box)
print()    

运行结果：可以看到识别到的文本

3.2 本地图像识别

用我自己手写的文字来试下，有标点符号，甚至还特别写了一个看不清的哎呀，让它识别一下

#OCR文本识别-本地
from azure.cognitiveservices.vision.computervision import ComputerVisionClient
from azure.cognitiveservices.vision.computervision.models import OperationStatusCodes
from azure.cognitiveservices.vision.computervision.models import VisualFeatureTypes
from msrest.authentication import CognitiveServicesCredentials
import time
import os
computervision_client = ComputerVisionClient(You endpoint, 
                                             CognitiveServicesCredentials( Your key))
local_image_path = os.path.join("C:\\Users\\Uniigym3\\AppData\\Roaming\\Python\\Python38\\Scripts\\images", "ocrTest2.jpg")
local_image = open(local_image_path, "rb")
read_response  = computervision_client.read_in_stream(local_image,  raw=True)
read_operation_location = read_response.headers["Operation-Location"]
operation_id = read_operation_location.split("/")[-1]
while True:
    read_result = computervision_client.get_read_result(operation_id)
    if read_result.status.lower () not in ['notstarted', 'running']:
        break
    print ('Waiting for result...')
print(read_result)    
if read_result.status == OperationStatusCodes.succeeded:
    for text_result in read_result.analyze_result.read_results:
       for line in text_result.lines:
           print(line.text)
           print(line.bounding_box)
print()    

运行结果：太感动了哈，它竟然识别出来了，甚至perfect的.都识别出来了 ，很有意思

 我尝试把照片倒过来，然后就识别不到那个不清楚的字了。